How to choose optimal settings Decision taking Mechanical Ventilation Patient Equipment Peter C. Rimensberger Pediatric and Neonatal ICU Department of Pediatrics University Hospital of Geneva Geneva, Switzerland

Defined Clinical Targets and Goals 1) Achieve good oxygenation and acceptable CO2 2) reduce WOB in spontaneous breathing patients limit peak pressure use lower Vt use higher PEEP 3) Try to protect the lung limit peak pressure use lower Vt use higher PEEP

ARDS network trial (6 vs.12 ml/kg) Small Vt ventilation ARDS network trial (6 vs.12 ml/kg) n = 861 Mortality: 31 vs. 38 (p < 0.007) PIP: 32 vs. 39 cmH2O Pplat: 25 vs. 33 cmH2O NEJM 2000;342:1301-1308

Have changes in ventilation practice improved outcomes ? Albuali WH Pediatr Crit Care Med 2007; 8:324 –330

Clinical variables associated with mortality Multivariate analysis Albuali WH PCCM 2007; 8:324 –330

Max. used Vt and effect on the developpment of ALI in ICU patients without lung injury 50 40 30 20 10 Mean Vt 10.9 ± 2.3 p < 0.001 n = 100 Proportion of ALI (%) n = 160 n = 66 < 9 9 to 12 > 12 Tidal Volume (ml/kg PDW) Gajic O et al. Crit Care Med 2004; 32:1817-1824 Physiologic Vt (normal lungs with spont. breathing) : 6 to 7 ml/kg

The concept of small Vt ventilation is a concept of “physiologic Vt ventilation” ARDS network trial (Vt 6 vs. 12 ml/kg) n = 861 Mortality: 31 vs. 38 (p < 0.007) NEJM 2000;342:1301-1308 Physiologic Vt (normal lungs with spont. breathing) : 6 to 7 ml/kg

… and in the neonate? … no RCT in newborn infants has substantiated so far the experimental finding that avoiding large tidal volumes … is lung protective in newborn infants.

NeoVent Ventilation practices in the neonatal intensive care unit: an international cross-sectional study Van Kaam A , Rimensberger PC (manuscript submitted)

NeoVent Ventilation practices in the neonatal intensive care unit: an international cross-sectional study Van Kaam A , Rimensberger PC (manuscript submitted)

The ARDS lung is rather small than stiff The baby lung The ARDS lung is small, with a normal aerated portion having the dimension of the lung of a 5- to 6- year old child (200 – 300 g of lung tissue as compared to 700 g) Gattinoni L Intensive Care Crit Dig 1987; 6:1-4 Gattinoni L et al. J Thorac Imaging 1988; 3:59-64 = percentage of the expected normal lung volume The ARDS lung is rather small than stiff

1) Adult and child: Acute respiratory distress syndrome (ARDS) ARDS is a heterogeneous lung disease 2) Neonate: (Infant) Respiratory distress syndrome (iRDS) iRDS is a heterogeneous lung disease

MRI signal intensity from non-dependent to dependent regions The water burden of the lung makes the lung of the preterm infant, despite surfactant treatment, vulnerable to VILI 4-day-old, 26-week gestation infant 2-day-old, 38-week gestation infant Adams EW AJRCCM 2002; 166:397–402

The baby lung. The ARDS lung is small, with a normal The baby lung The ARDS lung is small, with a normal portion having the dimension of the lung of a 5- to 6- year old child (200 – 300 g of lung tissue as compared to 700 g) Gattinoni L, Pesenti A Intensive Care Crit Dig 1987; 6:1-4 Airway pressure (cmH2O) Volume (l) The normal lung The baby lung Vt / kg ratio Vt / “baby lung” ratio Overdistention

 “permissive hypercapnia”, HFOV, ECMO or ECCO2-R Allowable Vt depends on pathology and disease severity The normal lung Volume (l) or pathologies with reduced TLC: - Lung hypoplasia CDH iRDS - Lobar Collapse - Lobar Pneumonia The baby lung Airway pressure (cmH2O) Vt of 6 ml/kg bw in a patient with a by 50% reduced TLC corresponds to at Vt of 12 “ml/kg”, he should therefore receive only 3 ml/kg bw !  “permissive hypercapnia”, HFOV, ECMO or ECCO2-R

ARDS network trial (6 vs.12 ml/kg) Higher PEEP during small Vt ventilation or peak pressure limitation ARDS network trial (6 vs.12 ml/kg) n = 861 Mortality: 31 vs. 38 (p < 0.007) PIP: 32 vs. 39 cmH2O Pplat: 25 vs. 33 cmH2O NEJM 2000;342:1301-1308 Oxygenation target

PEEP and FiO2 allowances in PEEP studies ARDS Network 6 versus 12 ml/kg: NEJM 2000;342:1301-1308 ALVEOLI: NEJM 2004;351:327-336 LOVES: Meade MO 2008;299(6):637-645

CT-aeration At ZEEP and 2 PEEP levels = turning up the PEEP approach poorly areated normal CT-aeration At ZEEP and 2 PEEP levels = turning up the PEEP approach Diffuse CT-attenuations Focal CT-attenuations Rouby JJ AJRCCM 2002;165:1182-6

“Anatomical” Recruitment Recruit to TLC (?) Gattinoni L AJRCCM 2001; 164:1701 Focus is on “opening” (re-aerating) previously collapsed lung units

The PEEP step approach: “Functional” Recruitment 25/10 Overinflation ends Pressure control ventilation PEEP 15 PEEP 20 PEEP 10 PEEP 25 40/25 Over- inflation starts Rimensberger 2000 (unpublished) Focus is on “opening”, but certainly on avoiding overdistending lung units P/F-ratio, oxygen delivery and quasi-static Crs during PEEP steps Lichtwarck-Aschoff M AJRCCM 2000; 182:2125-32

O2-improvement = Shunt improvement = a) recruitment PaO2 PaCO2 VA b) flow diversion PaO2 VA PaCO2 Gattinoni L (2003)

Prevalent overinflation = dead space effect PEEP 15 1 2 1 – PEEP PaO2 and PaCO2 increase Gattinoni L (2003)

PEEP titration: O2 and CO2 response Steps of 5 cmH2O to 40/25 Pressure control ventilation 25/10 25/10 Overinflation ends PEEP 25 PEEP 20 PEEP 15 PEEP 10 Overinflation starts

Understanding lung opening and closing Behavior of the whole lung: Hysteresis Behavior of a single alveolus Radford: in Respiratory Physiology (eds. Rahn and Fenn) 20 19

Lung opening and closing Frequency distribution of opening and closing pressure in patients with ARDS Crotti S AJRCCM 2001;164: 131–140 Behavior of the whole lung: Hysteresis Volume derecruitment throughout deflation UIPdefl Pclosing Alveolar recruitment throughout inflation LIP UIPinfl Radford: in Respiratory Physiology (eds. Rahn and Fenn) 20 19

Rimensberger PC Crit Care Med 1999; 27:1946-52 40 Pressure (cmH2O) Volume Intrigued by this basic physiologic observation we wanted to know what was now happening in such a lung that exhibits hysteresis during mechanical ventilation. First we generated in a rabbit with surfactant depleted lungs a quasi-static pressure-volume by by stepwise insufflation and desufflation of fixed volume aliquots. Then we repeated this maneuver imitating mechanical ventilation going up to 15, 20, or 25 cmH2O respectively and desufflated from this different pressure points (corresponding to the peak pressure during ventilation). You can observe, that by increasing the maximal insufflation pressure volumes increased on the volume axes with a deflation limb moving upwards, which means recruitment of lung volumes. In a next step we did a similar thing but going first to 3o cmH2O, corresponding to total lung capacity) and desufflated to 15, 10, and 5 cmH2O, repsectively, fllowed by a reinsufflation from these various pressure points that correspond to PEEP during ventilation. Pressure (cmH2O) Rimensberger PC Crit Care Med 1999; 27:1946-52

small tidal volume ventilation (5 ml/kg) 30 Pression Optimal PEEP Recruited vol 8 Lung recruitment allows to place the respiratory cycle on the deflation limb And this again allows then to explore the pressure-volume relationship of the lung to position the dynamic cycle in almost every place desired. Using small VT and a PEEP of 8 in this situation first, followed by a short increase to 30 cmH2O of airway pressures as a sustained inflation and reducing then again the PEEP to 8 as previously led to a gain on lung volumes with the tidal cycle now on the deflation limb. Note that we are now on a more compliant part of the curve. Rimensberger PC Crit Care Med 1999; 27:1946-52

Oxygenation response in two groups; with and without recruitment (identical PEEP) optimal-PEEP recruited vol. Gradient PaO2/FiO2 This maneuver improves oxygenation... Rimensberger PC Crit Care Med 1999; 27:1946-52 25

Lung recruitment: The optimal least PEEP approach optimal-PEEP recruited vol. Rimensberger PC Crit Care Med 1999; 27:1946-52 Rimensberger PC Crit Care Med 1999; 27:1940-5 25

This slide illustrates the effect of a single recruitment maneuver on lung volumes in a baby on ECMO for congenital diaphragmatic hernia.

The open lung concept searches for maintaining lung volume before RM after RM Halter JM AJRCCM 2003, 167:1620-6 alveoli per field inspiration expiration I – E PEEP before and after Rimensberger PC Crit Care Med 1999; 27:1946

Keep PEEP after RM above lung closing PEEP is an expiratory phenomenon to maintain the lung open Keep PEEP after RM above lung closing Rimensberger PC Crit Care Med 1999; 27:1946-52 Lapinsky SE Intensive Care Med 1999; 25: 1297-1301

Optimal = “Maximum dynamic compliance and best oxygenation at the least pressure required” So we could develop a strategy to position the tidal cycle optimally in the safe window, at the point of maximum dynamic compliance, at the least pressure required, which will result in good oxygentaton. This is illustrated in this mathematical lung model from Keith Hickling Hickling KG et al. AJRCCM 2001; 163:69-78

Volume Pressure TLC UIP CCP LIP Courtesy from David Tingay Lachmann and colleagues through their ‘OPEN LUNG CONCEPT’ work have described a technique in which ventilation can be applied near the top of the deflation limb of the pressure volume relationship. Lachmann utilises the relationship between lung volume and oxygenation to achieve this. Their technique involves incremental increases in mean airway pressure MOUSE CLICK, which recruits lung volume, until the best oxygenation is achieved. They have shown that this point should be close to Total Lung Capacity and any further increases in mean airway pressure will result in overdistension of the lung and impairment of oxygenation. As the lung is now recruited and because the lung displays hysteresis if mean airway pressure is decreased any resultant change in lung volume should occur near the deflation limb. Initially this will result in small changes in lung volume until the critical closing pressure of the lung is reached and subsequently any further decrease in mean airway pressure will result in rapid loss of lung volume, with resultant drop in oxygenation. Pressure Courtesy from David Tingay Royal Children’s Hospital Melbourne 38

Use of dynamic compliance for open lung positive end-expiratory pressure titration in an experimental study F Suarez-Sipman Crit Care Med 2007; 35:214–221

Get the lung as much homogeneous as possible Frerichs I et al. J Appl Physiol 2002; 93: 660–666 Volume distribution Frerichs I, Dargaville P, Rimensberger PC Intensive Care Med 2003; 29:2312-6

Volume distribution Tidal volume distribution Frerichs I, Dargaville P, Rimensberger PC Intensive Care Med 2003

Regional «homogeneity» on the deflation limb right lung non-dependent region normal lung injured lung post surfactant lung right lung dependent region normal lung injured lung post surfactant lung Dargaville P, Frerichs I, Rimensberger PC (submitted)

 similar time constants ARDS / RDS lung (Heterogeneous) Alveolar Rupture! C = 2 Vt=3ml/kg  similar time constants homogeneous Vt distribution Normal lung / recruited lung at optimal lung volumes Vt=6ml/kg C = 2 Vt=5ml/kg Vt=6ml/kg C = 1 VT= 1ml/kg  various time constants Very compliant healthy lung adjacent to noncomliant ARDS lung depicted by thickened wall  heterogeneous Vt distribution

Heterogeneous: Injured Lung Contrast this to the heterogeneously injured lung. On alveoli is collapsed and there is loss of interdependece. The remaining alveoli fill the void as they over- expand which is no longer buffered by the 4th alveoli. Unstable hetero lung. Instability and overexpansion (atelectrauma and volutrauma) all due to hetero lung ventilation Alveolar Overdistension into the Area of the Collapsed Alveoli

Courtesy from G. Niemann What we think we see here is in vivo evidence supporting this hypothesis. There is a collapsed alveoli and the neighboring alveoli bulges into the void. All the other alveoli have neighbors which prevent them from overexpansion Courtesy from G. Niemann

Homogeneous: Normal Lung Large circle is lung and smaller circles alveoli. Notice no way for overexpansion at peak inspiration d/t interdependence b/w alveoli ie each prevents the others from overexpanding The adjacent alveoli support each other to prevent overexpansion This is example of homo lung Minimal Change in Alveolar Size with Ventilation

Correlation of Inflection Points with Individual Alveolar R/D DiRocco et al. Intensive Care Med

Courtesy from G. Niemann What we think we see here is in vivo evidence supporting this hypothesis. There is a collapsed alveoli and the neighboring alveoli bulges into the void. All the other alveoli have neighbors which prevent them from overexpansion Courtesy from G. Niemann

Best approach to recruitment: «Open the lung and keep it open» Use the smallest Vt you can afford (you deal with a baby lung !) then you have to work you through to find the optimal least PEEP approach = “Functional Approach to Recruitment” Your tools at bedside: P/F ratio, PaCO2 and Cdyn

But there is a significant risk of overdistending many patients There is no sound rational for fixed PEEP and FiO2 schemes as used in PEEP studies ! ARDS Network 6 versus 12 ml/kg: NEJM 2000;342:1301-1308 ALVEOLI: NEJM 2004;351:327-336 LOVES: Meade MO 2008;299(6):637-645 But there is a significant risk of overdistending many patients